Objective. This proposal has the important goal of furthering our understanding of the longitudinal course of autism spectrum disorder (ASD). It is motivated by the urgent need for early prognostic markers able to explain the extensive heterogeneity of ASD outcomes. To this end, we propose a longitudinal study of young children with ASD (2-3 years) to identify the neurobiological underpinnings of early developmental changes in restricted repetitive behavior/interests (RRB) ? one of the most clinically impairing aspects of ASD - and their predictive contribution to later function. The proposal builds on the confluence of related findings and our own work including: 1) clinical evidence that RRB changes over the age window between ~29 and 42 months are prognostic markers of ASD adult functioning; and 2) advances in brain developmental functional connectomics that allow investigations of neural circuits in preschoolers with ASD using natural sleep MRI. We specifically aim to test 1) whether changes from 24-36 to 36-48 months-of-age (T1, T2) in the intrinsic functional connectivity (iFC) of somatomotor (SM) striatal-cortical circuitry are associated with changes in the repetitive sensory motor (RSM) subdomain of RRB; and 2) whether these early brain-behavioral changes predict later (age: 48-60 months, T3) adaptive functioning. Exploratory aims will examine the potential contributions of structural connectivity changes in striatal-cortical tracts, and test whole-brain iFC employing unbiased connectome-wide association. Finally, we will explore the value of an alternative, data-driven hierarchical clustering approach to characterizing outcomes based upon multiple clinical dimensions at T3. Methods. We anticipate obtaining complete T1 and T2 brain-behavioral data from 100 preschoolers with ASD enrolled at age 24-36 months and followed prospectively on a yearly basis. At T1 and T2, preschoolers will undergo natural sleep imaging with state-of-the-art MRI (high resolution T1- and T2-weighted structural MRI, multiband resting state fMRI (R- fMRI), and when possible, diffusion tensor) and phenotypic assessments rigorously selected to deeply phenotype a range of ASD core and associated symptoms. To examine the predictive value of iFC and RSM changes (i.e., T1-T2) to later function, children will be re-evaluated at 48-60 months (T3). A partial list of assessments includes: Autism Diagnostic Observation Schedule-2, Behavioral Observation Social Communication Checklist, Repetitive Behavior Scale-Revised, clinician and parent measures of comorbid psychopathology and adaptive functioning. Brain-behavior analyses will primarily rely on R-fMRI, and explore diffusion tensor imaging. Significance. Findings will elucidate the neural correlates of changes in RSM at the earliest practical time following clinical diagnosis. They will provide a developmentally informed understanding of the neural underpinnings of outcomes, thus bringing the field closer to a neural stratification of individuals. Such knowledge is essential for developing neuroscientifically-informed treatments. Impact is maximized by sharing de-identified data with the NDAR and ABIDE yearly to accelerate scientific progress.